Topic: Any way to increase FSB accuracy?
OK, I know its probably a common question but i am seeing some huge discrepancies between FSBs generation/backtests and MT4 backtests/live running.
I have an EA developed on the M30. Its really basic with no indicators for entry other than the default slot. Obviously FSB says its pretty profitable. Running a backtest in MT4 over the same period with 'control points' also shows its profitable and looks similar to the FSB result. Using 'every tick' (and yes i know its interpolated) in MT4 over the same period suggests the strategy is basically junk. So i let it run for a few days and then had a look at the MT4 backtest results vs what actually happened. The 'every tick' backtest is almost a perfect match for the real time run while the 'control points' is a lot more optimistic. I havent recalculated it in FSB but to be honest there seems little point. The original FSB generation was close to the MT4 control points backtest which i now know to be worthless.
So is there any way to increase the accuracy of FSB because it looks like when its generating strategies its doing a really loose version of MT4s 'control points' backtest. Is it perhaps that certain indicators/open methods are more susceptible to inaccuracy so just need to be avoided? Its not just one strategy, its a few but they are all based on things like :
'Enter at Upper/Lower band' on a channel type indicator
'Enter Long after a Downwards Move'
Or 'Previous High/Low', 'Top Bottom Price' type entries.
and all show similar where MT4 'control points' is similar to FSB and profitable but MT4 'Every Tick' is closer to live and not so profitable.